Motion adaptive 3D Y/C separation algorithm using motion estimation and motion compensation

This paper proposes an advanced motion adaptive three-dimensional (3-D) Y/C separation algorithm for NTSC signals. The basic principle of Y/C separation processing uses the fact that the phase information in a composite signal inverses every horizontal pixel, vertical line, and temporal field/frame. As such, this phase inversion characteristic is used to separate the Y and C components from a composite signal through an addition and subtraction operation. However, inaccurate separation introduces deterioration in quality, such as cross color and cross luminance, due to lack of regard for motion information. Accordingly, to acquire accurate movement information, the proposed algorithm uses a 2-D separated Y signal for detecting moving regions and estimates a motion vector using BMA-based motion estimation. Thereafter, the Y and C components are separated using the phase inversion characteristic and motion compensation (MC). Experimental results demonstrated a significant reduction in cross color and cross luminance in moving pattern regions.